Gender and Long-Term Care in Ghana: Women, Caregiving and Inequality Across the Life Course

Long-term care has a gender dimension long before an older person needs help with bathing, mobility, medication or everyday activities. It begins in the way unpaid work is distributed across households, in who reduces employment to support relatives, in who accumulates pension rights, in patterns of widowhood and longevity, and in the economic consequences carried from working age into later life.

These connections are increasingly important within the Ghana Ageing, Long-Term Care & Community Support Knowledge Hub. Ghana's 2021 census recorded almost two million people aged 60 and above, with women representing 56.7% of that population. The National Ageing Policy has long recognised gender disparities affecting older people and the need for gender-sensitive responses. At the same time, unpaid domestic and care work remains distributed unevenly between women and men.

The implications extend beyond today's caregivers. A woman who spends years combining informal employment with unpaid care may reach later life with fewer savings and weaker contributory pension protection. She may then experience disability or frailty herself while depending on another generation of women for support. Care inequality can therefore reproduce itself across the life course.

Ghana's strategic challenge is not to diminish family care, which remains socially and practically important. It is to make visible who provides it, what it costs, when it becomes unsustainable and how formal services, social protection and community infrastructure can distribute care more fairly.

Caregiving is part of Ghana's gender economy

Care is economic activity even when no payment changes hands. Preparing meals for an older relative, accompanying somebody to healthcare appointments, assisting with personal care, supervising a person with dementia and remaining available because somebody cannot safely be left alone all consume time.

That time has alternatives. It could have been used for paid employment, trading, education, rest or other family responsibilities. The fact that unpaid care frequently occurs within loving family relationships does not eliminate its economic consequences.

Available international gender data for Ghana illustrate the wider imbalance: women and girls aged 15 and above spend substantially more of their time on unpaid care and domestic work than men. Not all of that activity is elder care, but the disparity provides important context for a country in which population ageing will increase the number of households managing frailty, disability and chronic illness.

Gender therefore needs to be considered within family partnership and caregiver support rather than treated as a separate equality issue. If a long-term care strategy assumes that families will continue providing more care without asking who within those families performs it, the system can unintentionally convert demographic pressure into additional unpaid work for women.

This does not mean men do not provide care. Husbands care for wives, sons support parents and male relatives contribute financially and practically. Family arrangements vary substantially. The analytical point is that care responsibilities are not distributed evenly, and policy needs to understand the pattern rather than relying on a gender-neutral idea of “the family”.

The cost of care accumulates across the life course

A life-course perspective changes how long-term care inequality is understood. The person providing care at 45 may become the person needing care at 75.

Employment history matters because Ghana's retirement-income system includes contributory pension arrangements in which entitlement reflects participation and contributions. Many people also work in the informal economy, where employment and retirement security can operate very differently from continuous formal-sector employment.

Care responsibilities can interact with those differences. A woman working in informal trading may have flexibility to accompany an older parent to hospital, but every day away from trading can reduce household income. Another woman in formal employment may have stronger income security but less control over her working hours. A daughter living abroad may contribute financially while a sister living locally provides most direct care.

Over many years, repeated reductions in paid work can affect savings, business development, career progression and retirement resources. The immediate care arrangement may appear to cost the public system very little while generating substantial private costs that emerge later.

This is why fair work and responsible employment have relevance to long-term care. Employment practices cannot solve family caregiving pressures on their own, but greater recognition of caregiving can help prevent workers from facing an unnecessary choice between maintaining income and supporting somebody important to them.

Scenario: a daughter becomes the invisible financing mechanism

A 48-year-old woman in Kumasi operates a small trading business. Her 77-year-old mother has arthritis, hypertension and increasing difficulty walking. Two siblings contribute money, but because the daughter lives closest she becomes the person who attends healthcare appointments, collects medicines, buys food and responds when her mother is unwell.

No single task initially appears overwhelming. The cumulative effect is different. Her trading hours become less predictable, customers begin going elsewhere and she increasingly uses business income to pay transport and household expenses for her mother. When her mother falls, the daughter closes the business for several days.

The family is providing effective support, but the distribution of responsibility is unequal. Financial contributions from siblings do not compensate fully for lost working time, and the arrangement contains no contingency if the daughter's business fails or she becomes ill.

A stronger response would begin by making the care workload visible. Which tasks genuinely require the daughter? Could rehabilitation or mobility equipment reduce dependency? Could siblings fund practical assistance rather than only general household costs? Are healthcare appointments being coordinated efficiently? What would happen if needs increased?

The objective is not to monetise every act of family care. It is to recognise that unpaid labour is part of the care system's real resource base. Unless that contribution is understood, the apparent affordability of family-based long-term care can conceal economic costs carried disproportionately by one person.

Older women experience ageing from a different economic starting point

Gender inequality affects not only who provides care but who reaches later life with financial protection. Ghana's older population contains more women than men, and women are more likely to experience widowhood because of differences in longevity and marital patterns.

Later-life income can reflect decades of labour-market participation. Continuous formal employment can build contributory pension entitlement. Informal employment, interrupted employment and unpaid household work may provide less predictable retirement security.

This matters because long-term care is not financed solely through a dedicated national care entitlement. Families combine their own labour, income, pensions, social protection, healthcare coverage and privately purchased support. An older person's personal income therefore influences how much choice they have within that mixed system.

A woman with limited independent income may technically have family support while possessing little control over how money is spent. Another may own property but depend on relatives to manage transactions. Widowhood can alter household income and decision-making at the same time that health needs are increasing.

Gender-sensitive ageing policy therefore needs to consider income security, property, housing and decision-making alongside healthcare. Financial independence is not separate from person-centred care; it can determine whether someone can choose where they live, obtain transport, purchase assistance or refuse an arrangement they do not want.

Informal employment complicates the relationship between work, pensions and care

Ghana's labour market makes simple assumptions about retirement difficult. Census evidence shows that a substantial proportion of older people continue working, with most employed older people operating in the private informal sector. Work in later life may represent choice and continued participation, financial necessity or a combination of both.

Women may reach older age after decades of trading, agriculture, household production, caregiving and other forms of work that do not resemble a conventional formal employment career. Their economic contribution can be considerable even where contributory pension accumulation is limited.

That distinction matters for care planning. A 67-year-old woman who continues trading may also support grandchildren and care for an older husband. Describing her simply as an “older person” obscures several simultaneous roles: worker, caregiver, income generator and potentially future care recipient.

Policy therefore needs to avoid assuming a clean transition from employment to retirement followed by dependency. Ghana's care system will increasingly support people whose lives move between paid work, unpaid care and their own changing health needs.

Organisations exploring the interaction between workforce capacity, continuity and dependency can use the Predictive Workforce Risk Module to structure similar questions. Although it is not a Ghana-specific policy instrument, the principle applies equally to family systems: concentrated dependency on one individual creates continuity risk.

Widowhood can combine emotional loss with practical vulnerability

Widowhood is not a uniform experience. Many widowed women remain economically active, socially connected and independent. Others may experience a sudden combination of bereavement, income change, housing insecurity and increased reliance on adult children.

The distinction between vulnerability and dependency is important. A widowed older woman should not be presumed incapable simply because she lives alone. Ghana's census identified hundreds of thousands of older people living alone, and independent living can reflect preference rather than abandonment.

Risk increases where living alone intersects with poverty, mobility limitations, poor health or weak social networks. Gender can intensify those intersections because older women may have fewer financial resources and longer periods of later life to navigate.

A proportionate system therefore asks what the person can do, what support she wants and what is changing. Strengths-based support is useful precisely because it avoids treating widowhood, age or poverty as a complete description of somebody's capability.

Community relationships, religious participation, neighbours, savings, property, family contacts and personal skills can all contribute to resilience. Formal support should strengthen those assets without using them as justification for withholding help where substantial needs exist.

Caregiving can affect women's health as well as income

Long-term caregiving has physical and emotional consequences. Supporting somebody with significant mobility needs can involve lifting, transfers and interrupted sleep. Dementia can require continuous supervision. Coordinating healthcare, household responsibilities and finances creates cognitive as well as physical workload.

These pressures do not mean family care is inherently harmful. Many caregivers value the relationship and regard support as part of reciprocity between generations. The risk arises when commitment is interpreted as unlimited capacity.

A caregiver may continue providing assistance after developing back pain, hypertension, depression or exhaustion herself. Families can normalise increasing workload gradually because no single moment appears to mark the transition from manageable support to unsustainable care.

Caregiver health should therefore become part of prevention. This does not require medicalising family relationships. It means recognising that the health of the caregiver influences the safety and continuity of the person receiving support.

Prevention and early intervention should include the household around the older person where appropriate. A mobility aid that reduces lifting, rehabilitation that restores independence or occasional practical assistance may protect two people's health simultaneously.

Scenario: an older woman is both caregiver and person at risk of needing care

A 68-year-old woman in the Central Region cares for her 73-year-old husband following a stroke. She assists him with dressing, bathing and transfers and manages most household responsibilities. Their adult children contribute financially but live elsewhere.

She describes herself as managing well and initially declines outside help. During a healthcare contact, however, it becomes clear that she has developed persistent back pain and has stopped attending appointments for her own hypertension because leaving her husband is difficult.

The immediate temptation is to focus only on the husband's stroke. A wider assessment shows that the household has two health risks rather than one. If the wife's health deteriorates, both may require substantially more support.

Practical intervention could include rehabilitation for the husband, safer transfer techniques, appropriate equipment, clearer medication arrangements and agreement with family members about regular periods when somebody else is present. The wife's preference to remain her husband's principal caregiver can still be respected.

The governance lesson is that caregiver wellbeing should be visible before a crisis. A system that records only the identified patient's clinical condition can miss the fragility of the care arrangement surrounding that person.

The scenario also demonstrates why gender-sensitive care does not mean assuming women should stop caregiving. It means ensuring that their willingness to care is not mistaken for evidence that unlimited care can be provided safely.

Care expectations can influence daughters, daughters-in-law and granddaughters differently

Family caregiving is shaped by relationships as well as gender. Expectations may fall differently on daughters, daughters-in-law, spouses, nieces and granddaughters. Geography, income and family dynamics determine who actually provides support.

This can produce complicated negotiations. A daughter living overseas may finance care and believe she is carrying the greatest cost. A daughter living locally may feel that daily responsibility is invisible. A daughter-in-law may provide personal care because she shares the household despite not being the older person's closest relative.

These arrangements can work well when responsibilities are discussed and agreed. They become more fragile when care is allocated through assumption.

Person-centred practice therefore needs a family-centred dimension without allowing family preference to override the older person's own voice. Involving families and advocates is most effective when roles are clear: who provides practical support, who contributes financially, who communicates with services and who should be contacted when circumstances change.

Care planning should also avoid treating the nearest woman as the default caregiver. Asking an adult daughter what she is able and willing to do is different from assuming availability because she lives nearby.

Dementia can intensify gendered caregiving pressure

Dementia makes many of these issues more pronounced because the need for support can extend beyond identifiable physical tasks. A person may remain mobile while requiring supervision, reassurance, help with money, support with medication and protection from getting lost.

This form of care can be difficult to combine with employment because the workload is defined by availability rather than a fixed number of tasks.

Women caring for parents or spouses may therefore reduce paid work before the person requires substantial personal care. Behavioural changes can also affect relationships, sleep and caregiver wellbeing.

Ghana's developing dementia response needs to recognise families as partners while ensuring that family members receive information and support. Diagnosis without a practical post-diagnostic pathway can transfer responsibility back to households without increasing their capability.

This is particularly important where stigma limits discussion of cognitive impairment. A caregiver may conceal difficulties until a serious incident occurs. Accessible information, community awareness and links between healthcare and local support can make earlier help more acceptable.

The gender issue is not that dementia caregiving is exclusively women's work. It is that existing inequalities in unpaid care can become amplified when a condition requires prolonged supervision.

Rural and urban women experience different care constraints

Gender interacts with geography. In rural areas, distance from health and specialist services can increase the time required to accompany an older relative to appointments. Transport costs and limited service availability can make family assistance particularly important.

A rural caregiver may combine care with agriculture, informal work and household responsibilities. Adult children may have migrated to urban areas, leaving one relative with a disproportionate share of direct support.

Urban caregivers face different pressures. Services may be closer geographically, but congestion, employment schedules, housing costs and long travel times within cities can make practical care difficult. Paid assistance may be more available while remaining unaffordable to many households.

The same national policy can therefore produce different operational realities. Gender-sensitive long-term care needs to understand how health inequalities and prevention interact with location rather than assuming women's caregiving experiences are uniform across Ghana.

Metropolitan, municipal and district planning can help make these differences visible. Local evidence should identify not only the number of older people but household composition, disability, poverty, transport barriers and the availability of family and community support.

Scenario: migration redistributes care within a rural family

An 82-year-old woman lives in a rural community in northern Ghana. Three of her adult children have moved to cities for work. A fourth daughter remains nearby and combines farming, childcare and support for her mother, whose eyesight and mobility are deteriorating.

The siblings send money when possible. From their perspective, responsibility is shared. From the local daughter's perspective, almost every practical task remains hers: preparing meals, arranging transport, collecting medicines and responding at night.

The family begins discussing whether their mother should move to Accra to live with one son. She strongly prefers to remain in her community, where she has longstanding social and cultural connections.

A person-centred response does not assume either that the daughter should continue indefinitely or that relocation is inevitable. It asks which needs could be reduced or redistributed. Could visual assessment, mobility support or home adaptations improve independence? Could family contributions pay for regular practical help? Could visits be organised more predictably? What community resources are available?

If the older woman's needs later become too complex for the arrangement, that should trigger a new decision rather than being treated as failure by the daughter.

The scenario shows how migration can redistribute care without changing the family's sense of collective responsibility. Long-term care policy needs to understand the difference between shared responsibility in principle and shared workload in practice.

Paid care work can reproduce the same gender inequalities as unpaid care

Formalising long-term care does not automatically resolve gender inequality. Care work itself is often female-dominated internationally, and Ghana's emerging care economy may develop similar patterns.

If poorly paid or insecure work simply replaces unpaid female caregiving with low-paid female caregiving, part of the inequality has moved rather than disappeared.

Professionalisation therefore matters for gender as well as quality. Training, clear roles, supervision, employment standards and career progression can help establish care work as skilled economic activity rather than an extension of domestic labour.

This is especially important where household workers gradually take on personal care responsibilities. A worker employed initially for cleaning or cooking may begin assisting an older person with bathing, transfers or medication as needs increase. The change can occur informally without corresponding training, pay or oversight.

Ghana's future workforce strategy should recognise local employment and skills development as part of care-system infrastructure. Expanding community care could create employment, but employment quality will influence whether the workforce is sustainable.

Care workers also have their own families. Scheduling systems that rely on extreme flexibility can transfer care conflicts from service users' households to workers' households. Workforce design therefore needs to consider continuity for the person receiving support and reasonable employment conditions for the worker providing it.

Economic protection needs to extend beyond paying caregivers

Recognising unpaid care often leads quickly to a debate about caregiver payments. Financial recognition may be one option within some future policy designs, but economic protection is broader than a cash benefit.

Families can be supported by reducing the amount of avoidable unpaid labour required, protecting caregivers' connection to employment and making essential services more accessible.

Potential components include:

• practical training that makes care safer and less time-consuming;
• rehabilitation and equipment that preserve the older person's independence;
• respite or replacement support where continuous supervision is required;
• employment practices that recognise significant caregiving responsibilities;
• social protection that reduces extreme financial vulnerability; and
• greater access to affordable formal support as care needs become more intensive.

No single mechanism removes the economic cost of care. The stronger objective is to prevent that cost from being concentrated invisibly on one family member.

For organisations examining how community investment and workforce practices generate wider benefits, the Adult Social Care Social Value Report Builder provides a way to structure consideration of employment, community participation and social outcomes. Its measures require adaptation to Ghana's context, but the principle of recognising value beyond direct service activity is relevant.

Financial independence can be a safeguarding issue

Gender, ageing and financial security also intersect with safeguarding. An older woman who depends entirely on relatives for money may have limited ability to challenge an unsafe living arrangement or obtain assistance independently. Widowhood, cognitive impairment and reliance on somebody else to manage digital payments can increase vulnerability.

This should not lead to suspicion of ordinary family financial support. Intergenerational sharing is an important part of many Ghanaian households. The safeguarding question concerns whether the older person retains meaningful choice, whether money and property are managed with consent and whether dependency is being exploited.

Safeguarding, consent and human rights in later life therefore have a gender dimension. Protection from financial or property-related harm matters alongside protection from physical abuse and neglect.

Economic empowerment earlier in life can itself become a protective factor in old age. Secure income, savings, property rights and access to financial information can increase the choices available when care needs develop.

Long-term care policy consequently intersects with wider gender policy. A care system cannot correct every inequality accumulated over decades, but it should avoid reinforcing those inequalities by making access to support dependent entirely on personal wealth or unpaid female labour.

Scenario: protecting choice when money and care are controlled together

A 76-year-old widow in Greater Accra owns her home and receives regular income, but increasing visual impairment means that her adult son now manages many payments and purchases on her behalf. He also arranges transport and accompanies her to healthcare appointments.

The arrangement began with her agreement and remains largely supportive. Over time, however, she becomes concerned that she no longer understands some transactions and says she wants greater involvement in decisions about household spending. Other relatives dismiss the concern because her son is “looking after everything”.

A person-centred response does not begin by assuming either abuse or incapacity. It begins with the older woman. What does she understand? What assistance does she want? Can financial information be made more accessible? Are transactions transparent? Does she want another trusted person involved?

If evidence of coercion or financial exploitation emerges, appropriate protection and escalation become necessary. If it does not, relatively simple changes may restore control while allowing the son to continue providing practical assistance.

The scenario illustrates why support and control should not become indistinguishable. Families often manage money because it is practical, but good care preserves the older person's authority wherever possible.

Organisations working through similarly complex balances between autonomy and protection can use the Positive Risk-Taking Planner to structure decisions about choice and proportionate safeguards. It does not determine Ghanaian legal rights, but it can help make the reasoning behind support decisions more explicit.

Better data can make hidden care visible

Gender-sensitive long-term care requires evidence that goes beyond counting service users. Ghana already has valuable census and social data, but future planning will benefit from a clearer understanding of who provides unpaid care, how much care they provide and what consequences follow.

National averages can conceal substantial variation. Caregiving differs by age, income, employment, disability, household composition and geography. Older women are not one homogeneous group, and neither are female caregivers.

Useful system intelligence would connect several questions:

• who is providing substantial unpaid care and to whom;
• whether caregiving is affecting employment, income or health;
• which households are supporting people with high levels of functional dependency;
• where caregivers have little replacement support;
• how gender influences access to pensions, income and paid services; and
• whether interventions improve outcomes for both the person receiving care and the caregiver.

This does not require creating an intrusive national register of family caregivers. Population surveys, service assessments, local evidence and programme evaluation can progressively improve visibility.

The distinction matters because what is not measured can appear free. Unpaid care may disappear from conventional expenditure accounts even while determining whether hospitals can discharge people, whether disabled adults remain at home and whether older people can continue living in their communities.

Stronger data and quality metrics should therefore include the sustainability of care arrangements rather than focusing solely on formal service activity.

Governance should test who carries the consequences of policy

Gender-sensitive governance is not achieved simply by adding women to a policy document as a vulnerable group. Decision-makers need to test how policies redistribute time, money and responsibility.

A community-care programme may appear efficient because it reduces institutional expenditure, but if it assumes unlimited family availability its costs may have shifted into households. A hospital-discharge pathway may achieve rapid clinical transition while requiring a daughter to leave employment. A digital service may improve efficiency while assuming that a younger female relative will manage technology for an older person.

These are governance questions because they concern the real consequences of system design.

The Governance Maturity Assessment can help organisations examine whether accountability, evidence and escalation are sufficiently developed around complex service decisions. It is not a Ghanaian governance standard, but its underlying discipline is relevant: leaders should understand not only whether a policy has been implemented but who is carrying the operational risk created by it.

At national level, the Ministry of Gender, Children and Social Protection has an obvious policy interest in the intersection between ageing, social protection and gender. Health-sector institutions, pension and social-protection bodies, Metropolitan, Municipal and District Assemblies, community organisations and emerging care providers hold different parts of the wider picture.

No single institution can remove gender inequality in long-term care. Governance needs instead to make the connections visible enough that one system does not unintentionally create pressure for another.

Men need to be part of the care transition

A gender analysis that focuses only on supporting women risks leaving the underlying distribution of care unchanged. Ghana's future long-term care strategy also needs to consider how men participate in caregiving.

This is not primarily about prescribing private family roles. It is about avoiding policy and service practices that automatically direct care information, training and responsibility toward female relatives.

A son accompanying his father to an appointment should be considered a potential caregiver in the same way as a daughter. Male spouses providing intensive care need access to caregiver information and support. Community communication can present caregiving as a shared family and social responsibility rather than an inherently female role.

Greater participation by men does not remove the need for formal services. Redistributing an excessive unpaid workload more evenly still leaves an excessive workload. The objective is both fairer sharing within families and a stronger care infrastructure around them.

This connects with equality and inclusion at a wider level. Long-term care systems influence social norms through the people they address, the roles they assume and the evidence they collect.

A gender-responsive care system would strengthen choice rather than prescribe family life

There is a risk that gender-responsive policy becomes overly prescriptive: assuming women are burdened, men are absent or particular family arrangements are preferable. Ghanaian households are too diverse for that approach.

The stronger model is one based on capability and choice.

An older woman should be able to continue caring for her spouse if that role matters to her without being left unsupported. A daughter should be able to contribute to her mother's care without being presumed available for unlimited personal assistance. A son should be recognised as a caregiver where he performs that role. An older person should be able to accept family help without surrendering control over money or decisions.

Services can support these choices by asking better questions. Who is providing care? What are they willing to do? What are they trained to do safely? What effect is care having on them? What does the older person want? What happens if the arrangement changes?

These questions turn gender analysis into practical care planning rather than abstract policy language.

International learning lies in recognising the full care economy

Countries with more developed formal long-term care systems have adopted different mechanisms to support families, including public home care, caregiver benefits, respite, social insurance and employment protections. Those arrangements depend on fiscal capacity, administrative systems and labour-market conditions that cannot simply be transferred to Ghana.

The transferable principle is broader: unpaid care should be recognised as part of the care economy even when it remains outside formal expenditure.

Ghana also offers an important perspective to countries where professional systems have sometimes treated families mainly as visitors or informal supplements. Family networks can provide continuity, cultural understanding, advocacy and relationships that formal services cannot reproduce.

The challenge is therefore not formalisation for its own sake. It is finding a sustainable boundary between family relationships and care-system responsibility.

As Ghana's population ages, that boundary will increasingly determine gender outcomes. If formal support develops slowly while dependency increases, households will absorb much of the difference. Existing patterns suggest that women are likely to carry a substantial share. If community services, rehabilitation, caregiver support and a skilled care workforce expand alongside ageing, families can remain central without being expected to function as an unlimited substitute for infrastructure.

Preparing now can prevent today's care inequality becoming tomorrow's ageing inequality

The long-term nature of the issue creates an opportunity. Ghana does not need to wait until population ageing is substantially more advanced before considering gender within care-system development.

Workforce strategy can professionalise care occupations before low pay and insecure conditions become entrenched. Social-protection development can consider how interrupted and informal working lives affect later-life security. Community-care models can assess caregiver capacity explicitly. Health services can identify when a household arrangement is becoming unsafe. Data systems can distinguish between family presence and sustainable caregiving.

Most importantly, prevention can preserve independence. Every improvement that enables an older person to manage safely with less assistance reduces pressure on the person themselves and on whoever would otherwise provide that assistance.

Gender-responsive long-term care is therefore not a specialist programme operating at the edge of ageing policy. It is a way of testing whether the entire system distributes responsibility fairly and sustainably across people's lives.

Conclusion

Gender shapes Ghana's long-term care system through two connected realities. Women form the majority of the country's older population, while women and girls continue to undertake a disproportionate share of unpaid domestic and care work. The effects accumulate across the life course through employment, income, health, pension participation, widowhood and the availability of support in later life.

Family care will remain an important strength within Ghanaian communities, but its sustainability cannot be judged simply by whether relatives exist. The practical questions are who provides the care, how intensive it has become, what the caregiver is giving up, whether the older person retains choice and what happens when the arrangement can no longer continue.

A stronger future system would make those realities visible without undermining family relationships. Rehabilitation, community support, respite, better caregiver information, professional care services, fair employment and stronger social protection can each reduce the concentration of responsibility on individual women. Greater participation by men can improve the distribution of family care, although redistribution within households cannot substitute for adequate care infrastructure.

For Ghana, the opportunity is to address gender while the country's long-term care architecture is still developing. Doing so would connect ageing policy with economic security, workforce development and human rights rather than treating care as a private household matter. The result would not be the replacement of family care, but a more sustainable settlement in which caring for another person does not unnecessarily determine a woman's economic security, health or choices in her own later life.